High Performance Computing Resources
Center for Computational Omics and Research for Translational EXcellence is Meharry’s advanced on-premise data center. This HPC infrastructure enables high throughput computation, supports cutting edge AI and machine learning model development, and powers secure integration and analysis of large-scale data. This center provides essential computing architecture to meet the demands of medical scientific research and data intensive applications.
CORTEX features four HPC clusters: [flesh these out with more details]
- Axon: First Meharry supercomputer (Ada).
- Dendron: Quartet supercomputer via NSF MRI funding.
- Cognos: CPU-intensive Dell HPC for large-scale analytics via HEERF funding.
- Synapse: GPU-intensive Dell HPC for AI model training (HRSA-funded).
Key Features
- Computing Power
The Ada/Midframe leverages parallel computing architecture to effectively handle computationally intensive tasks. - Scalability
The system offers scalability, allowing users to dynamically assign computing resources to fit their requirements. - Data Handling
The Ada/Midframe combines data management methods and optimized data analysis tools that enable users to swiftly process and analyze vast amounts of data.
Computing Infrastructure
Our high-performance network is comprised of the Ada and Midframe supercomputing clusters. Both clusters have a head node and four compute nodes. The operating system is Suse Linux Enterprise and InfiniBand serves as the network fabric, providing high-speed communication between compute nodes.
Ada Head Node Specifications
- 2 EPYC CPU (128 cores total)
- 4TB memory
- 2TB NVME Global File system
- 192TB Disks Space.
Specifications for each Ada Compute Node
- 1 EPYC CPU (64 cores total)
- 512GB local memory
- 0.8TB NVME Local File system
Midframe Head Node Specifications
- 2 EPYC CPU (128 cores total)
- 4TB memory
- 2TB NVME Global File system
- 192TB Disks Space.
Specifications for each Ada Compute Node
- 2 EPYC CPU (128 cores total
- 2TB local memory
- 0.8TB NVME Local File system
- 192TB Storage Space
Software Environment
- Operating System: Suse Linux Enterprise for High-Performance version 15-SP4
- Compiler Suite: gcc (SUSE Linux) 7.5.0
- Parallel Programming Models: OpenMP
- Job Scheduler: SLURM (to be deployed soon)
Other Software Packages
- 5.1 AI and Data Science Suite
- R / Rstudio
- PostgreSQL
- MIOpen Library
- Tensorflow 5.2 Bioinformatics Suite
- Pytorch
- BioPython
- Pysam
- Pymol
- Rosetta 5.3 Julia Suite
Resources Available to Users
The Ada/Midframe gives a comprehensive resource to support the computational needs of its faculty and students. These resources include:
- Researchers can submit batch jobs to the system.
- This system combines a high-speed interconnect to enable the communication and data transfer between compute nodes, supporting faster parallel computations.
- A distributed parallel file system ensures users have fast and reliable access to files during computations.
- This system offers a selection of software libraries and tools specifically tailored for medical and health sciences research.
- A dedicated support team is available to provide technical assistance, guidance and troubleshooting issues to users.